Since the beginning of my Ph.D. program in 2013, I have developed a research interest in cutting-edge computational methods for collecting and analysing large complex datasets such as social media data. My PhD project, completed in 2017 and published as a monograph entitled “Online communities and crowds in the rise of the Five Star Movement” by Palgrave (Bailo, 2020), applied innovative methods to collect cross-platform social media data and analyse them through the lenses of both social network analysis and quantitative text analysis (NLP).

Since completing my PhD and working as a postdoctoral fellow and lecturer, I have further developed my quantitative research skills by collaborating with several transdisciplinary research teams. These collaborations resulted in the following top-tier venues:

I have previously published collaborative research applying network analysis methods to social media data in Information, Communication & Society (Bailo & Vromen, 2017) and Policy Studies (Bailo, 2015), research using spatial analysis of demographic and electoral data in European Political Science (Abbondanza & Bailo, 2018), and research based on the statistical modelling of questionnaire data in Political Studies (Vromen et al., 2016).

My current and future research is grounded in two lines of research about the

  1. Emergence, interaction, and diffusion of political opinions online; and the

  2. Impacts of social media technologies and services on political participation and organisation.

This research agenda recognises the existence of two distinct layers of activities (see Figure below). The top layer is defined by Cloud platforms and their information and networking technologies, such as artificial intelligence systems, search engines or social networking sites. The bottom layer is defined by patterns and relations emerging from our society and connecting people and groups. These two layers communicate through an evolving range of interfaces and devices. The Cloud layer is progressively becoming the dominant place for the diffusion of socially and politically relevant information. Its technologies have been observed to promote and foster information disorders at an unprecedented scale. The Society layer is the domain where people connect and interact as political actors and is also where they struggle to influence critical epistemic outcomes, that is, decisions about knowledge affecting the broader society.

image

1. Emergence, interaction, and diffusion of political opinions online

In the first research line on the Cloud layer, I plan to continue using a strong transdisciplinary approach to research the emergence, interaction, and diffusion of political opinions online around contentious issues and on developing practical measures for the degree of information disorder that can be found online around different topics.

Our research team, which I was instrumental in creating in 2019, has successfully attracted external and internal grants, including from Facebook (now Meta) for USD 60,000 and from the Australian Department of Defence for AUD 500,000, which we initiated in 2023 and are currently administering. The team engages in state-of-the-art transdisciplinary research with experts in digital ethnography, computer science and political science and aims at developing both methodological innovations to enhance the capacity to collect and analyse complex human-generated data as well as empirical discoveries while, at the same time, addressing questions concerning the impact of the current information and communication infrastructure on politics and society.

Three research questions guide the development of this line of research.

  1. What opinions do emerge online, but also when and where?

  2. How do opinions interact with others through the Cloud and its online communities?

  3. What factors determine the speed and reach of the diffusion of opinions?

Answers to these questions will contribute to improving the functioning of social media platforms - a critical place for public opinion formation. In addition to providing a practical understanding to develop applications and content moderation policies that foster discussion quality over content engagement, it will improve our understanding of regulating social media platforms to protect vulnerable groups, institutions and democratic processes (e.g. elections).

Researching cloud platforms involves a number of methodological challenges:

Collecting representative data about social media interactions

Although large volumes of social media data are available, many restrictions due to business decisions and ethical safeguards challenge collecting representative data, especially capturing unfolding events.

Adopting and designing collaborative tools

The interdisciplinary collaboration around research data that is both very large and complex (e.g. temporal, multidimensional, multimedia) requires using effective interfaces for both humans and machines to interact qualitatively and quantitatively with the data. Multimodal access is critical for data triangulation, combining qualitative and quantitative (or computational) methods to the data.

We address these challenges through methodological innovations, including the design of the following:

  1. Long-term online experiments;

  2. Computational techniques and software for cross-platform data collection and data matching;

  3. Machine learning pipelines for multidimensional (i.e., topic and polarity) text classification integrating both qualitative research, crowd-sourced input and large language model input;

  4. Research software applications to collaboratively define and manipulate data ontologies for opinion mapping and assist with automatically classifying large-scale collections of texts.

In addition to attracting significant funds, this research line has already resulted in the publication of two peer-reviewed publications (Bailo et al., 2023; Kong et al., 2022).

Our current research work is concentrated on the following:

  • The design of a dashboard for the real-time monitoring and analysis of social media data (see Figure on the right).

  • The development of a definition for information disorder as a global measure of any information space and an approach to quantify and measure it.

  • The empirical detection of information operations on social media and the development of modelling techniques to understand, measure and forecast their effectiveness.

  • The write-up of a handbook for using Wikibase, an extension of the software running Wikipedia to work with semi-structured data, in transdisciplinary research projects involving large data streams.

image

2. Impacts of social media on political participation and organisation

In the second research line, which is methodologically and conceptually linked to the first, I will continue studying the impact of social media technologies and services on political participation and organisation. This research will build on my study of Italy’s Five Star Movement and the role played by the Internet in affording and shaping its emergence, from a blog in mid-2005 to national government in 2018.

Using big data and comparative analysis, I am interested in further investigating the effect of Internet-mediated participation on political stability. Because Internet technologies and services can contextually reduce the importance of traditional communicative and organisational infrastructures (and their gatekeepers), their diffusion can theoretically result in the systematic emergence of new competitive political actors such as parties, movements, or leaders. Because social media platforms and their recommender systems can contextually downgrade the quality of the public debate, they can also stir political discontent and facilitate its mobilisation.

Three research questions guide the development of this second line of research:

  1. Can "Internet capital" offset a lack of social capital in facilitating participation and organisation?

    • Does Internet capital facilitate the organisation of political discontent from the margin (low-social capital areas)?

  2. Do Internet technologies and services create more political instability by facilitating the emergence of new political actors (individuals or organisations)?

  3. Do social media negatively affect the quality of political deliberation? What factors, technologies or services are important in reducing the quality of online conversation?

This research line has already resulted in the publication of a monograph with Palgrave (Bailo, 2020) and is currently concentrated on the following:

  • The development of theoretical work on the concept of "political epistemology", explaining the diffusion of politically contentious knowledge on social media. This is expected to inform the design of social media applications that are resilient to information disorders and the politically motivated diffusion of misinformation.

  • The write-up of two articles on the Internet-based mobilisation networks of the Five Star Movement. Based on the analysis of geocoded event data and survey data, these articles explain the role of both traditional social capital networks and Internet-based networks in the activation of militants and voters.

  • A survey-based research on the reconfiguration of political engagement after the fall of the Islamic Republic of Afghanistan in August 2021. This project offers new insights into whether the sudden fall of democracy can reconfigure civic networks for political engagement and the role played by social media in the reconfiguration.